After stroke, two opportunities to improve the patient's outcome are 1) to preserve the non-infarcted tissue in the ischemic penumbra and 2) to reperfuse severely ischemic tissue without producing further injury. To help devise useful protection strategies, we propose to investigate the factors that contribute to ischemic brain injury and propose the following hypotheses. 1) In partially ischemic tissue, electrical activity ceases but neuronal metabolism continues. With low O2, glycolysis increases to maintain ATP levels, producing lactic acid which accumulates because flow is reduced. Since ATP production by glycolysis cannot fully compensate for oxidative phosphorylation, AMP and purine levels increase so that tissue adenylates are irreversibly lost, either enzymatically or through clearance by blood. Stimulation of glutamate receptors and Ca++ influx accelerate these metabolic processes and increase tissue injury. 2) Upon reperfusion of severely ischemic tissue, depletion of tissue adenylates, through clearance by blood or enzyme activity, and propagation of oxygen free radicals, initiated during severe ischemia, add to rapid shifts of intracellular pH, Ca++ and osmolarity to worsen tissue injury. To test these hypotheses, a basic set of measurements will be made in vitro using two brain slice models of ischemia. To model the ischemic penumbra, we will use well-oxygenated 1000micro thick hippocampal brain slices. To create a model of reperfusion injury, 500micro and 1000micro slices will be exposed transiently to anoxia, unmodified or with 0 Ca++, lactic acidosis or both, and then returned to normal buffer. For both models we will 1) assess tissue injury histologically, 2) measure brain slice glucose utilization with 14C-2-deoxyglucose, 3) measure Ca++ influx with 45CaCl2, and 4) measure tissue lactate, ATP, ADP, AMP, adenosine, inosine, hypoxanthine, xanthine and uric acid. 5) We will also measure tissue hydroxyl free radicals after transient anoxia. There are four Specific Aims. 1) To test the hypothesis that persistent neuronal metabolic activity contributes to tissue injury in the ischemic penumbra, we will study well-oxygenated 1000micro and 500micro hippocampal slices during hypothermia or in the presence of NMDA, AMPA or metabotropic glutamate receptor antagonists for up to 4 hours in vitro. 2) To test the hypothesis that loss of tissue adenylates contributes to injury of the ischemic penumbra, we will study 1000micro and 500micro thick hippocampal slices in the presence of adenosine, the adenosine deaminase inhibitor deoxycoformycin (DCF) or adenosine with CDF for up to 4 hours in vitro. 3A) To test the hypothesis that reperfusion increases tissue injury after severe ischemia, we will expose 500micro and 1000micro brain slices to varying periods of anoxia to define a period of anoxia that consistently injures 50% of CA1 neurons and then study the effect of cooling slices to 33 degrees C prior to reoxygenation to determine whether the duration of anoxia or the reperfusion plays the major role in tissue injury. 3B) To further test this hypothesis and to define the relative roles of anoxia, low extracellular Ca++ and acidosis during severe ischemia, we will transiently expose 500micro and 1000micro hippocampal brain slices to anoxia, alone or with 0 Ca++, pH 6.4 or both, and determine whether these buffer modifications worsen the outcome seen with unmodified anoxia. 4) To test the hypothesis that depletion of adenylates during severe ischemia contributes to tissue reperfusion injury, we will expose 500micro and 1000micro brain slices to anoxia simultaneously with adenosine, DCF or adenosine with DCF.